Comparative study

To complete the permeability study, the effects of different reinforcement architectures on permeability are studied in this section. In section 4.2.3.1, the behavior of tex 200 UD flax/paper reinforcement studied above (through DOE approach) is compared with other la- boratory reinforcements made in this thesis, while in section 4.2.3.2 behavior of tex 200 UD flax/paper reinforcement is compared with two commercial reinforcements including Flax- Tape© and a woven glass fabric.

4.2.3.1. Comparison with other laboratory-made flax fiber reinforcements

Figure 4-25 shows the studied reinforcements and the comparison strategy used in this subsection. While fiber volume fraction is kept constant at Vf = 35 % for the experiments of this plan, comparison of results allows for studying the effects of different parameters (men- tioned in Figure 4-25) on permeability. These parameters are qualitative ones and cannot be studied through a DOE approach.

The first comparison uses two types of flax/paper reinforcements: the Flax/paper tex 200 (FP-200) and Flax/paper tex 400 (FP-400). They differ in linear density (tex) of the yarn and spacing between yarns.They are designed to have equal surface densities, so FP-400 consist of 12 yarns per inch compared to 24 yarns per inch in FP-200 (see Equation 3-4). These two reinforcements are shown in Figure 4-26.

The second comparison concerns the effect of paper/flax cohesion (IBS) on the permea- bility. To evaluate this, the paper layer of FP-200 is very cautiously peeled off and stacked over flax layer for testing. This reinforcement corresponds to Separated Flax/Paper Tex 200 (SFP-200) preform.

The third comparison is made between two UD flax reinforcements (without paper). The UD flax tex 200 (UDF-200) consists in 24 yarns/in. tex 200 UD flax layer made using the winding machine of Figure 3-4. It is observed that after fabrication of FP-200, its flax layer (Figure 4-26a) becomes more uniform and homogeneous compared to UDF-200 (Figure 4- 27) which is directly got from the winding machine. To evaluate this homogenization effect, the paper layer of FP-200, is very cautiously peeled off such that the flax layer remains intact, and therefore the pressed-dried UD Flax Tex 200 (PDUDF-200) layer is acquired following this procedure.

A fourth comparison between FP-200 and PDUDF-200 having identical UD flax layer allows for evaluating the effect of paper layer on the permeability. Table 4-12 shows the surface densities of the studied reinforcements. Surface densities of hybrid reinforcements are acquired through weighting laboratory samples and dividing them by the sample’s area.

However, since UDF-200 was fragile and likely to be distorted during weighting, surface density of this reinforcement is calculated using Equation 3-4.

Figure 4-26. Texture of (a) FP-200, flax layer side, and (b) FP-400.

Figure 4-27. Texture of UDF-200. Table 4-12. Surface densities of reinforcements.

Reinforcement name Reinforcement type Surface density (g/m2₎

UDF-200 UD flax tex 200 189

FP-400 Flax/paper tex 400 221 ± 1.5

FP-200 Flax/paper tex 200 216 ± 2.2

PDUDF-200 Pressed, dried UD flax tex 200 175 ± 2.3

SFP-200 Separated Flax/Paper tex 200 213 ± 1.7

Permeability results are summarized in Table 4-13 and depicted in Figure 4-28. Higher average and variance of FP-400 in comparison with FP-200 is attributed to open channels between tex 400 yarns and the consequent nesting effect, as explained for the 16 yarns/in. tex 200 reinforcements in section 4.2.1.2 based on Figure 4-10.

A noticeable observation is that FP-200 and SFP-200 show the lowest average K1 values and also the most consistent K1 permeability behavior considering that their standard devia- tion are much smaller than the other reinforcements. Comparing the results of these two re- inforcements also reveals that regardless of the paper bonding to the flax layer (magnitude of IBS = 0), the mean or variance of K1 permeability is not affected. The slightly smaller mean K1 in the case of SFP-200 could be simply due to variation of cavity height. During permeability tests for SFP-200, the cavity height was slightly reduced (around 0.05 mm) compared to permeability tests of FP-200 to provide more pressure on the SFP-200 preform and prevent any movement of the separated paper layers during injection, due to the effect of injection pressure.

Table 4-13. Results of permeability comparative study. Reinforcement type Nominal Vf (%) K1 (10-12 m2) K2 (10-12 m2) Ave. STD CV(%) Ave. STD CV(%) UDF-200 35 70.7 13.3 18.9 8.12 1.18 14.5 FP-400 35 77.1 22.5 29.2 7.64 1.34 17.6 FP-200 35 23.0 2.19 9.51 5.88 0.96 16.3 PDUDF-200 35 49.0 8.60 17.5 13.6 4.42 32.4 SFP-200 35 14.5 2.00 13.8 4.92 0.97 19.7

Figure 4-28. Results of permeability comparative study, (a) permeability along yarns (b) permeability perpendicular to yarns.

UDF-200 has shown around 1.5 time higher mean and standard deviation for K1 than PDUDF-200. As showed in Figure 4-27, the open spacing between yarns in the UDF-200 reinforcement and the consequent nesting effect are probably responsible for higher mean values and standard deviations of K1 of UDF-200 compared to PDUDF-200.

The PDUDF-200 reinforcement has shown two times higher mean K1 and four times higher K1 standard deviation than FP-200. From this results, it appears that paper layer in FP- 200 partly obstructs the open channels of flax layer and thus reduces the average K1. More- over, the presence of paper reduces the variability of permeability. This is probably a conse- quence of the homogeneity, uniformity and reduced nesting effect in the FP-200 reinforce- ment caused by the presence of paper compared to PDUDF-200.

In terms of K2 permeability, PDUDF-200 have shown higher average values than the other reinforcements. However, PDUDF-200 has also shown highest variability. Other reinforce- ments are more or less at the same order for both mean and variance of K2.

Practical experience during preparation and testing of PDUDF-200 indicates that this flax layer becomes very fragile after peeling off the paper ply from FP-200 reinforcement, even more than initial UDF-200 flax layer. Figure 4-29 shows two typical defects in this preform during permeability testing. Uneven flow front in Figure 4-29a is due to distortion of preform during preparation for permeability test and yarn separation in Figure 4-29b could happen at high injection pressure of permeability test, thus increasing variability in the results.

Figure 4-29. Defects in PDUD-200 reinforcement, (a) uneven flow front and (b) yarns sep- aration during permeability test.

From a permeability perspective, the reinforcements developed in this thesis represent a major improvement compared to the first generation of flax/paper reinforcements reported in a previous study [71], where the K1 and K2 permeabilities for a tex 1000 UD flax layer (with 820 g/m2 _{surface density) were respectively reported 4.98 ± 1.34 × 10}−12 _(m2_{) and 1.77 ±} 0.318 × 10−12 (m2), much lower than those in Figure 4-28 for FP-200.

4.2.3.2. Comparison with commercial reinforcements

Permeability behavior of hybrid flax/paper reinforcement (FP-200) is compared with two commercial reinforcements: a commercial UD flax reinforcement, FlaxTape© 200 (FT-200) supplied by Lineo Inc. (France) with surface density of 200 g/m2, and a plain weave glass fiber fabric reported in [60] and called Syncoglas R420 in this reference. FP-200 and FT-200 differ mainly with respect to texture of the fibrous network. Flax fibers on FT-200 have al- most no twist and therefore the reinforcement is more hairy than FP-200 as observed by comparing Figures 4-30a and 4-26a. Syncoglas R420 is shown in Figure 4-30b and its surface

density is reported 420 g/m2. The gaps between warp and fill yarns are also reported Gw = 0.58 mm and Gf = 0.35 mm, respectively. Figure 4-31 also shows the comparison strategy of this subsection.

Figure 4-30. Texture of (a) FT-200 and (b) Syncoglas R420.

Figure 4-31. Plan of permeability comparative study with commercial reinforcements.

Results of permeability tests are summarized in Table 4-14 and depicted in Figure 4-32. The warp (K1) and weft (K2) permeabilities of Syncoglas R420 are reported K1 = 1.79±0.398×10-10 and K2 = 1.43±0.301×10-10 respectively, at a fiber volume fraction Vf = 41.7 % [60]. These values are respectively one and two orders of magnitude higher than the

FP-200. It is believed that lower average K1 of FP-200 is mainly due to different architectures of the two reinforcements, with more flow channels in the case of Syncoglas R420, and also due to inherent hairy and rougher surface of natural fibers reinforcements compared to syn- thetic ones [66, 67].

In the case of average K2, the much lower performance of FP-200 is mostly related to the unidirectional configuration of the flax yarns in the FP-200. Higher standard deviation of Syncoglas R420 is due to spacing between yarns and nesting effect. Permeability of a twill weave flax fabric at Vf = 35 % (measured with engine oil) is reported around 0.4×10-10 m2 [72]. This is in the same order of magnitude than flax fiber reinforcements of this work and close to K1 permeability of the 16 yarns/in. reinforcements in Table 4-3.

K1 permeability of FT-200 is higher than FP-200 in terms of both mean and standard deviation and is very similar to that of UDF-200, reported earlier. This similarity between FT-200 and UDF-200 could be due to their quite similar fiber configuration consisting of longitudinal flax fibers which provide tiny longitudinal channels for fluid flow, while not having paper layer in both cases. However, FT-200 is made of long untwisted strand placed side by side while FP-200 is made of low twist yarns placed side by side.

In the following section some mechanical properties will also be investigated to determine whether the new reinforcement’s global characteristics show potential industrial applications.

Table 4-14. Permeability comparison with commercial reinforcements. Reinforcement type Nominal

Vf (%) K1 (10-12 m2) K2 (10-12 m2) Ave. STD CV (%) Ave. STD CV (%) FP-200 35 23.0 2.19 9.51 5.88 0.96 16.3 FT- 200 35 69.7 16.8 24.1 13.0 1.30 10.0 Syncoglas R420 [60] 41.7 179 39.8 22.2 143 30.1 21.05

Figure 4-32. Results of permeability comparative study, (a) permeability along yarns/warp direction (b) permeability perpendicular to the yarns/weft direction.